Opinion Classification System Using Supervised Learning Algorithm

Swati Ambasta, Surbhi Bhatia

Abstract: Due to huge amount of data posted online, decision making process considering opinions play a crucial role in everyone's life. Analysis in the field of making decisions and setting policies has shown that sentiment analysis and Opinion mining lies at the intersection of Question Answering system and Computational Linguistics. World Wide Web has tremendous amount of unstructured data present in web forums, social networking sites and other social platforms as reviews which diverts our study towards mining the opinions on web. Our research work focuses on extracting tweets, classifying them into positive, negative and neutral category and finally providing a recommendation as whether to buy or reject a product. Classification system is been proposed by using twitter data using Naive Bayes algorithm and accuracy of the evaluation strategies by has been evaluated. Review data is collected for various product domains from micro blogging sites like twitter, face book.

Keywords: Opinions Mining, Twitter, Sentiment Analysis, Naive Bayes